It is not possible to orthorectify a data file in ENVI if it does not have associated RPCs (rational polynomial coefficients).  So, what you need to do is first build the RPCs (via Map ->Build RPCs) and I am afraid this is going to require that you select GCPs.  You don't need to go to the field to do this, but you will need a reliable reference image to select GCPs.   The information you have is not enough to place the image on the earth accurately.  A landsat image should work if you believe the referencing to be reliable.

We also have the IMU data from the airplane for each photo. (Pitch, Roll, Yaw). Does this make any difference?

You will need the omega, phi and kappa values which can be calculated from the IMU.  They are non-trivial calculations so I wonder if your data provider has a means to calculate them for you.

Hi again. I have tried several times using this method and it seems to be out about 80m. The arial photo company used a gyroscope when taking the photos. By using a gyroscope, would it essentially make the omega and phi = 0.0?   Thanks

From what I understand, the gyro should be mounted on the sensor so the roll, pitch and yaw should be the residual rotation that was not able to be overcome by the gyroscope.  But how well the sensor is calibrated can affect the results as well.  For example, was the sensor properly calibrated when it was mounted on the plane?  A colleague of mine says they used to calibrate during flight a couple of times because the sensor can be shifted during flight.  So the resulting offsets you see could be a lack of proper calibration. After you build the RPCs, can you geographically link the image to a reference image to see how good the RPCs are?  I may be necessary to select at least a few (1-3) GPCs during orthorectification to improve the positional accuracy.